Something tells me I could be doing much better because I spend many hours a day just to learn 60 cards. Between the first review and the last, including breaks, maybe 7 hours will have passed. I've read comments where people say their retention is ~90% and they only spend 2-3 hours a day learning 45 cards (WHAT?!).

I've tried several approaches. I'll explain what happens in the anki stats image below.

1. In the beginning, I studied just 20 new cards/day mixed in with reviews all on Anki's standard settings. The amount of cards I had to "relearn" seems surprising now.

2. I took a day off, then experimented with cramming forgotten cards on Custom Study. The next day, you can see virtually no red (relearned cards). I felt pretty happy that day. The next day, however, wasn't so good. I forgot a lot of the young cards and attributed my previously much higher recall to the cramming I did.

3. 60 cards/day begins (alternating 20 new cards with forgotten cards within the last 1-3 days). Lots of forgotten cards again, it does not feel good to be sick of forgetting. In fact, I find it infinitely better to be sick of seeing a kanji than failing to recall it. I took a day off because it didn't seem to work.

4. Started back up, thinking I'd have another crappy day, my recall was phenomenal. I did it again and again and again and again and again, with a little drop in performance because I chose to SRS the lot before going to bed once. The next day, despite doing that, I could still recall >90% of the kanji and thought I'd give cramming a break (wondering if I'd been wasting time). Then it happened. Worst day of my SRS life.



Is there some kind of magic method people are using to study just 2-3 hours a day and maintain a level of retention that I need twice as long to keep up with? Or, maybe, my memory is below average(?) D:

Edit: also should point out that the cards I'd "over-review" weren't forgotten that day, only a huge portion of the new cards from the day before.

Remember that people that boast on the site are probably doing much better than the average learner, and yes, some of us just have terrible memories.

I spent roughly 2 hours a day to do my minimum 10 cards and my reviews, and then could do maybe ten more cards an hour, but my retention rate was fine, 85% or so. I think I have a below-average memory, but, I didn't come into RTK as a novice to Japanese so my numbers are all skewed anyway.

More important than how much time you spend learning/reviewing/cramming is -how- you're learning. Do you have the original book? Are you doing the visualizations as he suggests? Are you writing out the kanji with pen and paper when you learn them? When you review them? Are you making up your own stories, or just using the ones from the book or the top ones from the site? Have you tried a vocabulary-list style learning exercise? Are you breaking your study up into timeboxed periods?

Your study -method- is much more important than your study -time-, at least in this case where you're putting in -plenty- of time.

I do have the original book but I stopped using its visualisations to help me memorise after I decided to start cramming. Now, I only use visualisations on kanji where they can be simple. For example, "deceased eye" comes to mind when the card for "blind" comes up. Sometimes I create more abstract visualisations but that's not too often. One example is I remember violence by associating it w/ a derogatory/racist statement (it's a "white person's direction", pretty weird but it works for me).

Yes I write out every kanji before clicking answer (big notepad and pen). If I forget, I write it after.

I think giving a story to each kanji (to remember) only slows me down. I prefer the feeling of cramming because while I still have to employ some mental effort to remember, it's significantly less when there's no associated story, and just a couple related words (or rotely memorised).

SomeCallMeChris Wrote:Do you have the original book? Are you doing the visualizations as he suggests?

This was really important for me and not often stressed enough.

Animosophy, I had the exact issues as you until I read and re-read the two paragraphs on the visualization process. The more I closed my eyes and imagined the situation in the story turning into a kanji the more success I had.

I also found more success with kanji→keyword and felt this was a more natural transition to Japanese text. I have a feeling that some people will learn better from keyword→kanji and others from kanji→keyword. (Basically depending on your left/right brain learning preference you'll find more success with one than the other.)

Oh, well. If you're brute forcing with just a list of element names and no story, then 65% is what you should expect for a retention rate until repetition starts to make up the difference. If you want less repetition, you should do the stories. They don't have to be -his- stories (which run out pretty soon anyway), but his stories, site stories, your own stories, whatever.

If -that- doesn't work you can ask about alternative mnemonic ideas for people with a poor visual imagination, but without any mnemonic you're about par.

It seems that I'm more of a simultaneous processor.

I'll have another read through RTK and consider spending more time on visualisation, and maybe cut down on my cramming by a third, then compare my daily retention rate. It's not that I don't like spending this much time on kanji, I just don't want there to be an equally effective method that takes half the time.

SomeCallMeChris Wrote:If you want less repetition, you should do the stories. They don't have to be -his- stories (which run out pretty soon anyway), but his stories, site stories, your own stories, whatever.

If -that- doesn't work you can ask about alternative mnemonic ideas for people with a poor visual imagination, but without any mnemonic you're about par.

Would learning/making stories take comparatively less time than my current method?

1. 30-60 minutes on forgotten cards
2. 15-30 minute break
3. 20 new cards (~20 minutes)
4. 15-30 minute break
Repeat at least 3 times (60+ new kanji/day)

The bottom line: I'm doubtful whether visualisation would allow me to study 60 cards in one go and ditch the frequent reviewing (in order to stay above 90% retention). The one thing that prevents me from changing my study method is the fear of it changing for the worse (like it did recently).

The math is complicated, but, there's a serious time penalty for a low retention rate. Using a simpler-than-real doubling algorithm, you'd be looking at a perfectly retained card as
1d 2d 4d 8d 16d 32d 64d 128d 256d (512d ...)
9 reviews until you're looking at over a year between reviews, which is pretty much the same as 'totally learned' even if you're still reviewing kanji 3-4 years after starting RTK.

If you remember just a little better after every fail then you might have
1d 1d 2d 1d 2d 4d 1d 2d 4d 8d 1d 2d 4d 8d 16d 1d 2d 4d 8d 16d (32d... )
20 reviews before you've reach a 1-month interval which is 'reasonably well learned'.

Probably most cards won't be either extreme (and certainly not so nice and neat as my second pattern where you remember one interval longer each time).

It's too complex a math question to actually work out unless you're doing a PhD in this stuff, but it doesn't take a lot of failures before you're spending a -lot- more time reviewing than you would have if you'd just learned the item clearly in the first place.

And personally, I'm deeply demotivated by high failure rates, both because I find the increased repetition dull and frustrating and because I just don't like failing stuff. Only you know how your motivation is, but it sounds like you find high failure rates pretty discouraging. That mental angle is an important factor on its own - kill your motivation and it takes more effort to study less effectively and you might even make yourself unhappy to the point of quitting. Sometimes you have to go easier on yourself. 60 new characters a day is -a lot-. 20 is pretty speedy, 10 is a slow cruise... 60 is supersonic-jet speeds. Maybe you only need to be going at racecar speeds.

What I did
● Gradually learn the radicals through heisig
● use more than one definition
● read through stories and images that connect the radicals and kanji together.
● Get a stroke order image from jisho.org.
● Learn or browse through the most common words with the kanji to get a feeling for its usage
● make it simple




Try some of these ^ to help increase your retention rate.



An accurate description

Nest0r Wrote:Keep in mind that learning a particular meaning isn’t the intent of Heisig, it’s learning the writing and recognition of the kanji, with writing aiding recognition. Keywords are simply useful mental anchors for relational strategies. True focus on the meaning in kanji comes as meaning is invested through usage in words. There are what are called “compounding schemata” for the way kanji-based morphemes interact in compounds, and a morphological awareness of this is developed when learning the words and sentences.

http://darkjapanese.wordpress.com/2011/1...-and-kana/

Yes stories make a ridiculous difference, though adding will be slower at first and take you a while to get used to it you will save significant time on reviews. Especially in the beginning you'll need to spend extra effort making personalized stories and vivid images because (at least for me) it was something weird and unnatural.

After a few weeks of doing it though it becomes second nature, and basically any story will stick in your head to the point where you can add a hundred a day with minimal effort. Depending on how creative you are it could take more time, but once you get in the groove the stories'll add around ten seconds to your adds because they'll be second nature

Don't cram, unless you're taking it for a class and need to know those at that moment, it will just screw up your timing and cause extra work. I know it's hard, but you need to trust the SRS, that was probably the hardest part for me.

Also when people are talking about retention they typically aren't talking about short term retention but mature card retention. You only really need to start worrying if your mature card retention drops below 70 then you should probably take an adding break but you're doing just fine and if its around 90 you can take advantage of that to add more cards. Especially if you're adding lots of cards your short term retention will get pretty low. Looking at my stats my learning phase card recollection was around 65% and I managed to pull through just fine in a bit over a month.

Don't worry about the numbers, worry about if you're feeling overwhelmed, that's the best indicator of whether you need to change pace/method

Definitely do the stories! You obviously know about this site too, which makes it so easy to use them. You don't even have to think of stories yourself, just find the one on this site you like the most and then visualize it. It takes a couple of minutes extra per card tops, and it's going to make your reviewing much easier.

What is everyone talking about? Those are way good stats! I had 10 times the amount of that relearn constantly and I'm doing fine now.

[Zgarbas]

When you spend 7 hours cramming of course your retention rate will be low... it's too tiring.
Use the stories and visualize; otherwise there's no point in RTKing. Don't spend that much time studying (I mean, if you can afford studying for 7 hours go for it, but don't use them brute memorizing kanjis; variation is important).

If the visualization really doesn't work for you, then drop RTK and kanji-only studies and focus on vocab with kanjis on the side. Vocab will reinforce your kanjis and viceversa, so it will be less brute kanji memorization. You should already know enough to start out with a decent vocab, and just add kanjis as you encounter them in vocab. There are thousands of common words which require only kana and a few hundred kanjis.

The stats aren't that bad, and try to keep your own pace instead of focusing on other's; people have different learning styles. The more you worry the more frustrated you'll get and instead of studying you'll just burn yourself out till you drop Japanese; it's easy to go down that path.

Memory is a fickle whore at best, in my experience. I found myself maintaining above or around 80% on RTK and Core2000 decks, during daily reviews, even while adding anywhere between 60~100 new cards a day. You'd think this would imply a decent sort of overall memory, right? I struggle remembering dates (even days of the week, or what I did three days ago), and have, for the past few years, only remembered by birthday because other people reminded me.

You sort of have to find what works for you in respect to prioritizing things to remember. The stories on this website are probably a lot more relevant than the stories in RTK, and tend to get stuck in your head for a while until no longer needed. But maybe other aspects may be more important for yourself, such as assessing the environment you study in, diet, sleeping patterns, general head-space, etc.

Some people develop personal rituals (without being aware of it) which help them get into the right frame of mind in order to focus on something specific. I don't know you personally, but even something which may seem odd -if all else fails-, like breathing exercises before you do your reviews, or taking a few minutes break for every set amount of time spent revising, may help.

Stress can be a giant troll blocking the gates of memory. Maybe focusing less on the numbers and more on how you are feeling while revising, could help reduce stress. The numbers aren't really that important- treat Anki like a dance partner: you do your part and focus on the movements and Anki will arch her neck, sway her dress and keep your timing correct without you having to think about it.

tokyostyle Wrote:... imagined the situation in the story turning into a kanji...

Although I did try to visualise the stories, I clearly didn't do enough imagining of the actual primitive elements turning into the kanji. Now I'm finding it pretty diffcult to recognise a meaning (i.e. the keyword) in trying to learn to read. e.g. I can see that "Spiderman is doing something with a broom and a sow" - but how that could lead me to the meaning/keyword "affinity" is beyond me.

tokyostyle Wrote:I also found more success with kanji→keyword and felt this was a more natural transition to Japanese text. I have a feeling that some people will learn better from keyword→kanji and others from kanji→keyword. (Basically depending on your left/right brain learning preference you'll find more success with one than the other.)

I am a left-brain "successive processor". Could that be why I'm having trouble recognising kanji→keyword? Throughout RevTK I resisted reviewing in both directions, since Heisig doesn't recommend that. My understanding was that being able to recognise the keywords from the kanji would be a by-product of learning keyword→kanji, but that has not proved to be the case for me. So I'm now spending time having to review keyword→kanji. And I therefore would encourage the OP to take notice of the good advice given here to visualise the stories.

Aspiring Wrote:What I did
● Gradually learn the radicals through heisig
● use more than one definition

Do you mean more than one keyword? While I was doing RevTK 1, a single keyword to focus on was enough. But now that I'm learning to read, listing several meanings on the cards seems like a great idea. Thanks!

scarby dancer Wrote:I am a left-brain "successive processor". Could that be why I'm having trouble recognising kanji→keyword?

It may be more to do with the fact that the RTK methodology was created to be used in a purely keyword→kanji recognition. Heisig sort of explains why, but to put it one way, by doing RTK, kanji→keyword, you are sort of going against the flow of the method and making your brain use a different set of pathways and processes to activate the recall process.

I would make the comparison that the two are processing with data stored in individual regions, and having to perform kanji→keyword is sort of forcing the brain, under Heisig's RTK system, to use the keyword→kanji pathways in order to reach the kanji→keyword region, causing extra distance to travel and being interrupted by the keyword→kanji center on the way to the kanji→keyword region. Very generally speaking, our brains prefer to take the route of less resistance, which is why creating new connections takes time, and as RTK is based on creating temporary keyword→kanji association connections, it means this is less taxing than having to create an axillary region and using the keyword→kanji association has a pathway to access it.

If that makes any sense, at all.

SomeCallMeChris Wrote:And personally, I'm deeply demotivated by high failure rates, both because I find the increased repetition dull and frustrating and because I just don't like failing stuff. Only you know how your motivation is, but it sounds like you find high failure rates pretty discouraging. That mental angle is an important factor on its own - kill your motivation and it takes more effort to study less effectively and you might even make yourself unhappy to the point of quitting. Sometimes you have to go easier on yourself. 60 new characters a day is -a lot-. 20 is pretty speedy, 10 is a slow cruise... 60 is supersonic-jet speeds. Maybe you only need to be going at racecar speeds.

Exactly, it's very frustrating, very depressing. Anything below ~75% and I don't feel very good for the rest of the day. It's like you said, I just don't like failing stuff. Although I could drop down to 40, it's also frustrating when I'm not working on kanji/grammar in my free time (i.e. all day). 60 new cards takes up just enough time to keep me from wasting it on something less productive. On the flip side, it leaves me little time to study Tae Kim's grammar guide (but still enough to add 10 new sentences to my japanese-english deck). I want to prioritise grammar/sentences after I've gone through some ~3000 kanji.

Aspiring Wrote:What I did
● Gradually learn the radicals through heisig
● use more than one definition
● read through stories and images that connect the radicals and kanji together.
● Learn or browse through the most common words with the kanji to get a feeling for its usage
● make it simple

stratzvyda Wrote:After a few weeks of doing it though it becomes second nature, and basically any story will stick in your head to the point where you can add a hundred a day with minimal effort. Depending on how creative you are it could take more time, but once you get in the groove the stories'll add around ten seconds to your adds because they'll be second nature

AlgoRhythmic Wrote:Definitely do the stories! You obviously know about this site too, which makes it so easy to use them. You don't even have to think of stories yourself, just find the one on this site you like the most and then visualize it. It takes a couple of minutes extra per card tops, and it's going to make your reviewing much easier.

Zgarbas Wrote:Use the stories and visualize; otherwise there's no point in RTKing.

I haven't attempted it properly, so I know I'm at fault for not even trying... I'll do RTK like I'm supposed to and use the stories/visualisations for a week or so, or until I'm farmiliar with its use/purpose/effectiveness for me. Thanks for the tips on where to look, and I will definitely be adding more keywords per card like Aspiring.

stratzvyda Wrote:Don't cram, unless you're taking it for a class and need to know those at that moment, it will just screw up your timing and cause extra work.

I've noticed recently that despite instantly producing a kanji when shown its keyword, I'm sometimes very aware that it's because I saw it a lot the day before. This makes me wonder just how great an interval I can tolerate before forgetting again, IF I stop cramming. In situations like this I occassionally select "Hard" instead of "Good". Of course, over a significant amount of time, it's not like I'll forget everything when I taper off my custom review volume, but it does force me to remember that it's -spaced- repetition. In other words, my recall sometimes feels too "short-termish" (like I saw a kanji 5 seconds ago, even though it's been 12 or so hours), if that makes sense. Maybe it's irrelevant, so long as I remember?

Zgarbas Wrote:When you spend 7 hours cramming of course your retention rate will be low... it's too tiring.

No no, it wasn't the cramming, it was because I [i]didn't[i/] cram while going through 60 new kanji/day that caused the huge drop (68%) in recall. Plus, I actually find the hours I put in fairly comfortable. I know what you mean by the importance of variation, but the pace I've kept so far is what I'm most happy with. I would feel worse to reduce the new kanji I go through daily if I don't feel like I'm putting in equally valuable study time elsewhere (e.g. grammar, vocab). Kanji makes it easier to quantify my progress.

uisukii Wrote:Memory is a fickle whore at best, in my experience. I found myself maintaining above or around 80% on RTK and Core2000 decks, during daily reviews, even while adding anywhere between 60~100 new cards a day.

:|

Do you do the stories/visualisations? I spend many hours sitting, but I also dedicate about 6 hours a week exercising to keep my mind/body in decent shape. Right now I just want to maximise the economy of my study method.

Animosophy Wrote:

uisukii Wrote:Memory is a fickle whore at best, in my experience. I found myself maintaining above or around 80% on RTK and Core2000 decks, during daily reviews, even while adding anywhere between 60~100 new cards a day.

:|

Do you do the stories/visualisations? I spend many hours sitting, but I also dedicate about 6 hours a week exercising to keep my mind/body in decent shape. Right now I just want to maximise the economy of my study method.

Sounds like that is one of your problems, then. The more you try to "maximize" a human system, the more pressure and stress you inevitably place on it, which will in turn impact your retention performance (performance anxiety, anyone?). If trying to make things more efficient isn't working, try simplifying. Spaced repetition systems are constructed around a longer term time frame; consistency over time is more important than consistent minor tweaks for short term effect.

Find a way to have a sense of fun and reward your time with Japanese so it will be something you look forward to the next day. It's a little like football training: the kids who tend to perform at a higher level aren't because of "raw talent", or any arbitrary line, but because they look forward to turning up to training, compared to those falling behind who find it a time sink.

Positive attitude and enjoyment are the most "efficient" approaches to allow new information to flow into your brain with the least resistance.

I admit I'm concerned about my progress, and it is not an internal conflict. The single most rewarding feeling I get out of learning Japanese is moving forward. but I'm not so enthusiastic that I ignore my own mental capacity.

In hindsight, I was (and probably still am) hasty, but at least now, after re-reading the paragraphs on imaginative memory, I'm committed to using visualisations as the method intends.

I'll be reporting back on how much it helps me, if anyone's interested in the difference it makes (in my case, at least).

Thanks everyone for your help.

You guys really do make a helpful community

Well it's been 6 days. I said I'd come bk, so...



Studied 248 cards in 85 minutes today.
Again count: 11 (95.6% correct)
Learn: 0, Review: 237, Relearn: 11, Filtered: 0

Studied 255 cards in 93 minutes today.
Again count: 25 (90.2% correct)
Learn: 85, Review: 168, Relearn: 2, Filtered: 0

Studied 147 cards in 61 minutes today.
Again count: 12 (91.8% correct)
Learn: 0, Review: 135, Relearn: 12, Filtered: 0

Studied 189 cards in 63 minutes today.
Again count: 12 (93.7% correct)
Learn: 0, Review: 153, Relearn: 36, Filtered: 0

Studied 254 cards in 92 minutes today.
Again count: 13 (94.9% correct)
Learn: 9, Review: 206, Relearn: 39, Filtered: 0

It's working!

Instead of taking 7 hours to finish 60 new cards, I now do 100 new cards in ~3 hours, 20 mins then do my daily reviewing after. New goal is to finish the deck on March 30th or before April at the latest. Then I attack JP101's grammar bank. Seems Tae Kim's guide isn't as comprehensive as I thought. The huge yellow cram sesh was just to go through every kanji I rolely memorised before to attach a story/visualisation to it. This forum = awesome!

Animosophy Wrote:Is there some kind of magic method people are using to study just 2-3 hours a day and maintain a level of retention that I need twice as long to keep up with?

Yes. They're learning those words some place else. I would bet anything (strike that, I would bet a moderate amount of money) that your problem isn't bad methods or bad memory, it's simply that you're not as advanced a student as the people who have a better retention rate in Anki.

When I did Core sentences for instance, I found that I had encountered most of the words before (not the Kanji, but I heard the words in songs, subtitled shows, movies, anime etc. I've been watching for years). That lead to excellent progress. When I ran into words I've never seen before later on, trouble soon followed. But I fought my way through that too. I haven't looked at my retention rates, but I'm sure they went down.

Learning Japanese takes time. 2000 hours is the "expert" estimate, and that's only to "fluency", not mastery. If you've been doing 7 hours a day, you're doing amazingly well. Keep at it.

If anything, you could loosen up a bit, spend just 5 hours on hardcore study (including reasonably sized breaks). Take the other two hours and spend them watching something you love in Japanese (it's fine to watch it with subs, if you can't follow it without). Being two demanding of yourself can be counter-productive.

Edit: Oh, you're referring to RtK. I really should start reading full threads before commenting. Yes, the advice you already got was right. I won't delete my post though. Might help someone struggling with vocab. But none of what I wrote applies to RtK, except the part about you doing amazingly well by studying 7 hours/day. When it comes to learning anything, ass in the seat is always by far the main deciding factor.